The 1995 Survey at Black Assarca Island, Eritrea

DATE OF SURVEY: 28 May 1995

SURVEY TEAM: Mr. Ralph K. Pedersen, Dr. Chris Hillman, Mr. Yassin Adem

THE SITE

The site appeared to be a ship or boat wreck off the Assarca Islands (figure 1). It consisted of ceramics and ceramic sherds spread over an estimated area of seven by fifteen meters.  The center of the site was characterized by a small pile of broken amphoras and assorted sherds lying in disarray.  This grouping is hereafter referred to as the main pile (figure 2).   The site was discovered in 1995 by a group of snorkeling tourists led by Doi Malingri.   

The main pile sat on sand.  Scattered about, in an area approximately 5 by 7 meters, were several amphoras and numerous sherds.  These were found exposed, buried (figure 3), lying under coral shelves, and concreted into the coral (figure 4). 

Sherds and wood fragments were found down slope of the main pile for about another 10 meters.  This material lay mostly exposed.   None of the sherds examined showed fresh breakage and some were clean of marine growth on their underside.  This indicates the sherds may have been sitting on the slope for a long while.

The uppermost visible amphora lay exposed on top of the coral in approximately four meters of water (figure 5).  There was no evidence in the exposed materials as to either the ship's size or to the pattern of stowing the cargo of amphoras.

POTTERY TYPES

Four kinds of ceramics were observed.  These are called here Types 1 through 4.  Three were varieties of amphoras.  The nature of the fourth type remains undetermined.

       Type 1- This was a conical amphora with a short neck and a soft shoulder (figures 6 and 7).  The body tapered from the shoulder in a straight line and ended in a blunt toe, or button,   4.5 cm in diameter and 1.5 cm. high (figure 8).      The body was covered with broad, horizontal ridges spaced 1.2 cm. apart.  These gave the amphoras a corrugated appearance.  Due to marine growth, it was unclear whether these ridges continued to the top of the neck.

 There were at least 10 examples of Type 1 readily visible on the site.  While more examples were found buried, as well as concreted into the coral reef, it is not known whether these represented complete examples or fragments.

     All the free-lying Type 1 amphoras were broken.  Two, however, had intact bodies, missing only handles and neck.  One of these sat atop the main pile.  It was 71 cm. long and had a maximum circumference of 80 cm. at the shoulder.  The neck opening was 9 cm. in diameter.  The second intact body sat to the right (looking upslope) of the main pile.  Its length was 74 cm.

  Examination of a complete neck and handle assembly (figure 7) on one amphora with a broken body (to the right of the main pile) revealed the following dimensions:

     mouth diameter: 11.5 cm. outside, 9.5 cm. inside

     handle length: 12 cm.

     neck height: 13 cm. (from juncture of handle at shoulder)

     lip: 3 cm. high, 1 cm. thick.

     body fabric thickness: 1.5 cm. (approximate)

     Type 2- This was a rounded amphora, possibly flattened somewhat on one axis like a pilgrim flask.([1])  The body was covered with the same style of ridges found on Type 1 amphoras, but here they ran vertically.  All visible examples of this type were broken (figure 9).  The only apparent complete neck and handle assembly was raised by the wreck's discoverer (figures 11 and 12).  It is presently in the keeping of the Ministry of Marine Resources in Massawa.  The neck and handles visible on site were the same style as those on the raised sherd.  The neck of the raised sherd had at least five horizontal ridges.   The dimensions of this sherd were as follows:

     mouth diameter: 9.1 cm outside, 6.3 cm. inside the lip.

     lip: 1.2 cm. thick.

     throat at mouth and neck base: 4.5 cm. in diameter

     neck height: 11 cm.

     body thickness:1.3/1.4 cm.

     fabric color: reddish brown

     neck diameter at handles: 7.5 cm. (approximately)

     neck diameter at body joint: 11 cm.

Type 3- Only one example of this type was discovered (figure 12).  Found by Yassin Aden, it was in the coral under an overhanging outgrowth.  The neck, handles, and possibly shoulder, appeared to be missing.  The base was buried and was not examined.  The body had the same ridges as Type 1, and it appears to be conical.  This amphora, however, was broader than Type 1 with a diameter approximately 7 cm. greater than Type 1.

Type 4- Only one sherd of this type was found.  It was approximately 10 cm. long and 1 cm. thick.  There were no ridges or corrugations.  Unlike the red/brown fabrics of the other types, this was light brown in color.  There were no other discernable features.

Other artifacts/evidence: Two, possibly three, small wood fragments were found downslope of the main pile.  These were approximately 10 cm. long and 2 cm. thick and wide.  They were hard to the touch and appeared calcified by the coral.  The wood was dark in color.  There were no other distinguishing features.     No other possible hull features were seen.  It is not known whether the wood fragments were wreck material, or if they were associated with the remains of a Stalin's Organ lying nearby. No other artifacts, including anchors, were found despite the digging of several small test pits approximately 15 cm. deep to determine the extent of the wreck.  It is probable more artifacts lie under the sand, as well as concreted into the coral.

PRELIMINARY DATING

My original opinion of the date of the pottery was 7th century.  This was based on the Type 2 sherd raised by the sport diver.  I believed, however, a date a few centuries earlier or later was also possible.

ANALYSIS

Research has revealed that my initial dating was accurate.  The Type 1 amphora has close parallels from at least three sites in Eritrea and Ethiopia: Aksum, Matara, Adulis; from at least one site in Egypt: Elephantine Island; and from the shipwreck at Iskandil Burnu, Turkey.  Various examples have also been found around the Mediterranean Sea.

      The Aksum amphoras- (figure 13) These were excavated by the British Institute in East Africa in the 1970's by Neville Chittick.  The best-stratified examples date from the 5th to 7th centuries, although their dating is "notoriously difficult."[2]

     The Askum amphoras are similar to those from Assarca: "...the jar material is of the long, thin, conical corrugated surface type.  The base is generally buttoned.  The rim is gently inswept after its shoulders, with a straight lip with a diameter of 11-12 cm.  The rim is around 10 cm. long and carries a pair of opposed vertical loop handles."([3])

     Munro-Hay reports similar amphoras at Aksum with shop marks of northern Egyptian origin.[4]

      The Matara amphoras- (figure 14) Amphoras found at Matara by Francis Anfray are similar to those from the Assarca Island shipwreck.  Anfray states the ruins and objects at Matara apparently belong to the period from the 4th through 7th centuries.([5])

      The Adulis amphoras- (figure 15) These vessels in the National Museum in Asmara were excavated at Adulis, the port-city of the Aksumites.  Adulis fell to ruin sometime in the 7th century.  While the date of the city's demise is not known, its decline may have begun with an attack by Islamic forces in A.D. 640.([6])  The following year, the Byzantines surrendered Egypt to the Arabs, ending the Mediterranean connection with Aksum and the Indian Ocean.([7])  Adulis died sometime thereafter as the trade that was its lifeblood vanished.  Thus, the Adulis amphoras should date to the first half of the 7th century at the latest.

     Paribeni reports finding examples of conical amphoras at Adulis with marks on their bodies.  These marks, inscribed or painted on in red, were of Byzantine origin.([8])

       The Elephantine Island amphoras- (figure 16) Excavations on Elephantine Island on the Nile have revealed amphoras similar to those found in Eritrea and to the Assarca Island Types 1, 2, and 3. 

     Gempeler reports two types of conical corrugated amphoras: one with a slender body and narrow shoulders; the second with a broader body and stronger, wider shoulders, with a diameter up to 6 cm. greater than the narrow type.

     These amphoras date to the heyday of the Egyptian-Nubian wine trade, which falls between the years A.D. 550 and 750.([9])  Amphoras of varying form but with the same or similar corrugated appearance are also found at other Egyptian sites of the 6th and 7th centuries.([10])

        Gempeler also records pilgrim flasks at Elephantine, one of which (figure 17), while not an exact match, appears to have an affinity to Assarca Type 2.  This pilgrim flask, of unnoted size, has corrugations running vertically.  It does not, however, have ridges on the neck as does Assarca Type 2.  The Elephantine pilgrim flask dates to the 3rd or 4th centuries.([11])  Possibly, Assarca Type 2 represents a stylistically similar pilgrim flask (NOTE: Upon excavation in 1997, these types were found to be globular with only a slight flattening, if any.  RKP 2002).

     Iskandil Burnu- The shipwreck at Iskandil Burnu, Turkey contains amphoras similar to those found at Assarca.  Surveyed by INA over three years beginning in 1981, but never excavated, conical amphoras were found in association with several other types of pottery, including another amphora type besides the conical ones.  The amphoras were the main cargo of the ship.  It was hypothesized the ship had loaded the amphoras at Gaza or Ashkelon sometime in the late 6th century.  Both these cities were loading ports for Palestinian wine.([12])

     Other sites- Amphoras similar to the Assarca amphoras have been found at various places around the Mediterranean, such as Spain and Carthage, in strata datable from the mid-fourth century A.C. to the sixth century.([13])

     The dates for the conical, corrugated amphoras from all these sites, fall within the range from the 4th to 7th centuries, except for some that may belong to the 8th century.  Most probably, the ceramics from the Assarca Island shipwreck site date to these centuries.

CONCLUSION

The wreck at Assarca Island is a unique find.  From the evidence of other archaeological finds, the wreck appears to date from the 4th to 7th centuries after Christ.

     As far as we know, no other ancient wreck has been found in the southern Red Sea.  Indeed, with the possible exception of a Roman shipwreck at Zabargad Island, Egypt, it is the oldest shipwreck yet found in the entire Red Sea.

     At this time it is impossible to determine the nationality of the ship.  Nor is it possible to determine whether the site involves the remains of a ship headed to India, Arabia, or to some point on Africa's Indian Ocean coast.  Possibly, the vessel was a local trader carrying goods along the Eritrean coast or to as yet undiscovered Aksumite settlements in the Dahlak Archipelago.  In any case, the site holds great potential for our understanding of Red Sea commerce and seafaring in late antiquity, of which we know little. 

     Our knowledge of ancient maritime trade on the Red Sea relies in great part on classical authors such as Pliny the Younger and the anonymous author of the Periplus of the Erythraean Sea.  These writers recorded the kinds of cargoes carried by Red Sea ships during the period of the Roman/Byzantine Empire and the Kingdom of Aksum.  However, the information they gave is far from complete as the authors mentioned items they considered important.  Ships' cargoes often included contraband, private cargoes carried by individual crew members, personal belongings, and simply mundane items.  These cargoes generally were not recorded, and remain unknown.

     Ships also carried gear for repair and maintenance, fishing, cooking, and the weighing of cargo, as well as the personal items of the crew.  We have no record concerning any of these items for the ancient Red Sea, from either literature or excavations.  It is only from the scientific excavation of shipwrecks that such information can be gained.

     A shipwreck also can help refine the chronology of the past.  While a site on land can have many occupation levels, blurring the material record of a site, a shipwreck represents a discreet instance.  All the artifacts in a shipwreck come from a single horizon.  While a few artifacts may predate others by decades, or in rare cases even centuries, the bulk of a wreck's artifacts represent the materials and technology of one particular period.  As such, the excavation and study of ancient ships and cargoes can answer questions as can no terrestrial site.

     Indeed, as seen in the examples of amphoras from land sites in Eritrea and Egypt, dates for such sites often can only be determined to within a few centuries.  Thus, the chronologies of such sites and cultures are inadequate for our understanding.  On a shipwreck, however, there is greater accuracy due to the more-concise timeframe involved. 

     On many shipwreck sites coins are found.  In the case of the Assarca shipwreck, coins of Aksumite and Roman/Byzantine origin may be present.  We know from ancient literature that such coins were routinely carried on ships plying the Red Sea.  If coins are present, we should be able to discover the year of the ship's sinking.  In turn, we will be able to date accurately the ceramics, providing a chronology for sites on land where similar materials are found. 

     Thus, the shipwreck at Assarca is not only important for the history of Eritrea, but also for our understanding of all the archaeology of the Red Sea.

Ralph K. Pedersen, M.A.

Research Associate, Institute of Nautical Archaeology

 NOTES

          ([1]). The wreck's discoverer noted this type was flattened on one axis, yielding a shape like a fat disk.  None of the pieces of this type observed by the author displayed this feature.  Either the diver had an overactive imagination, or our survey team missed more intact examples.         

              ([2]). Munro-Hay, S.C., Excavations at Aksum: an account of research at the ancient Ethiopian capital directed in 1972-4 by the late Dr. Neville Chittick. BIEA, 1989 p. 314.

          ([3]). Munro-Hay, p. 314.

          ([4]). Munro-Hay, p. 314.

          ([5]). Anfray, F., Les Anciens Ethiopiens: siecles d'historie. Armand Colin Editeur, Paris 1990. p. 118.

          ([6]). Munro-Hay, S.C., Aksum: An African Civilization of Late Antiquity. Edinburgh University Press, 1991. p. 260.

          ([7]). Ostrogorsky, George, History of the Byzantine State. Rutgers University Press, 1969. p. 133.

          ([8]). Paribeni, R., "Richerche nel luogo dell'antica Adulis," in Monumenti Antichi, vol. XVIII, 1907, p. 551. 

          ([9]). Gempeler, Robert D., Elephantine X: Die Keramik Romischer bis Fruharabischer Zeit. Verlag Philipp von Zabern, Mainz am Rhein, 1992. p. 191.

          ([10]). Sauneron, S., Les ermitage chreitiene du desert d'Esna. Institut francais d'archeologie orientale du Caire, 1972. pp. 8,10.

          ([11]). Gempeler, p. 184.

          ([12]). Lloyd, M., "The Shipwreck at Iskandil Burnu." In INA Newsletter, 12:3, pp.4-5.

          ([13]) Keay, S.J., Late Roman Amphorae in the Western Mediterranean.  BAR International Studies 196, 1986. pp. 356, 358, 471.

For an account of the 1997 excavation see:

Ralph K. Pedersen, "Under the Erythraean Sea: An Ancient Shipwreck in Eritrea."  INA Quarterly, 27.2/3, Summer/Fall 2000.

Citation Information:

Ralph K. Pedersen

1996, Survey Report: The Shipwreck at Assarca Island, Eritrea.